def parse(self, fileName,entitySet,vectorMap):
#unpickling dict for cheap link
with open("entityDictionary.dat", "rb") as r:
entDict=pickle.load(r)
with open(fileName) as f:
for line in f:
if 'inv idx' in line:
break
elif 'predicate' in line:
predicate=self.extractPredicate(line)
currentVector = []
vectorMap.put(predicate,currentVector)
elif ': ' in line and 'num preds' not in line:
#if linking:extract single words
#if linking: link single words
#if linking: concatenate the entity links
origEntities=self.extractEntities(line)
l = Linker()
entities=l.cheapLink(origEntities, entDict)
count=self.extractCount(line)
index = entitySet.getIndex(entities)
pair = (index,count)
currentVector.append(pair)
print("done parsing")
return vectorMap, entitySet
